Billed by the minute · no subscriptions

Pay only for what you render.

Hourly billing. No subscriptions. You're billed for the exact time your machine runs, and billing stops as soon as you hit Stop.

Starter

RTX A4000

16 GB VRAM · Light rendering, ML inference

$0.25/hr
  • Rendering
  • Inference
  • Streaming
Deploy
Popular

Standard

RTX 4090

24 GB VRAM · Fast rendering, gaming

$0.49/hr
  • 4K Rendering
  • Gaming
  • ML Training
Deploy

Pro

L40S

48 GB VRAM · Production rendering, training

$0.99/hr
  • 3D Rendering
  • AI Training
  • Virtual Production
Deploy

Power

A100 80GB

80 GB VRAM · Large-scale AI training, HPC

$1.99/hr
  • LLM Training
  • HPC
  • Batch Inference
Deploy

All available GPUs

Live pricing from our infrastructure providers.

GPUTypeVRAMvCPUsRAMOffersPrice
RTX 4090Consumer24 GB212 GB20$0.49/hrDeploy →
RTX A4000Pro16 GB28 GB15$0.25/hrDeploy →
L40SDatacenter48 GB216 GB2$0.99/hrDeploy →
A100 80GBDatacenter80 GB832 GB7$1.99/hrDeploy →

Prices shown are GPU-only hourly rates. Additional costs for vCPUs, RAM, and storage apply.

RenderBox vs. buying hardware

Buy a GPU PCRenderBox
Upfront cost$2,500–$4,000$0
Monthly cost (5 hrs/week)$30–$60 electricity~$10
Upgrade cost$500–$2,000$0 — just pick a different GPU
PortabilityDesktop onlyAny device, anywhere
Setup timeBuild + configure: daysUnder 60 seconds

Common questions

When am I charged?
Only while your instance is running, for the exact time it ran — checked every 5 minutes, plus a final charge the moment you stop.
Any hidden fees?
None. GPU per-hour plus small costs for vCPUs, RAM, and storage. What you see is what you pay.
Can I switch GPUs?
Yes. Create multiple instances with different GPUs. Each has its own storage.
What happens if I run out of credit?
We'll send you a warning when your balance is low. If it reaches zero, your instance is stopped automatically. Your data is preserved.
Why do prices vary?
GPU pricing is dynamic based on availability and region. We show the lowest available price for each GPU model.

Ready to start?

Deploy your first render node in minutes.

Browse GPUs How it works